Objective To use a physics-informed neural network(PINN)-based model to predict hemodynamics in intracranial aneurysms and address the problems of long simulation time and high computational cost in traditional computational fluid dynamics(CFD)simulations.Methods The PINN model was trained using only the computational domain coordinates and sparse velocity measurement points from CFD data of clinical patients.The predicted blood flow velocity,pressure,and wall shear stress(WSS)from the PINN model were compared with CFD simulation results.Results The proposed method was used to test and validate data from four different patients.For velocity prediction,the average mean absolute error(MAE),average mean relative error(MRE),average mean squared error(MSE)was 4.60％,6.61％,and 0.229％,respectively.For WSS prediction,the average MAE,MRE and MSE was 5.54％,8.58％,and 0.510％,respectively.The PINN model demonstrated a good generalization capability across different aneurysm models and could reduce the computation time of hemodynamics from several hours to just a few seconds.Conclusions The PINN model can effectively compensate for incomplete measurement data through physical constraints,even when boundary conditions are unknown and measurement data are sparse.It can rapidly and accurately simulate the hemodynamics of intracranial aneurysms.This method has the potential to provide effective support for clinical risk prediction in intracranial aneurysms.

Objective To conduct preoperative simulations of three different osteotomy line design schemes under centric occlusion based on two distinct material assignment methods,evaluate biomechanical properties of the models,and explore which osteotomy line design schemes are more suitable for different types of mandibles.Methods Three types of mandibles were selected,and CT images were obtained for three-dimensional(3D)reconstruction.Material assignment was completed using the cortical/cancellous bone assignment method and the gray value assignment method.Osteotomy was simulated according to the three osteotomy line design schemes,followed by finite element analysis.Results In all simulation results of the mandibles,the maximum stress was 81.10 MPa,the maximum strain was 0.035 52,and the maximum displacement was 432.4 μm.The stress distributions obtained by the cortical/cancellous bone assignment method showed a larger stress distribution range than that that by the gray value assignment methods,but the maximum stress,strain,and displacement were generally lower.For the outflare type and common type mandibles,Scheme 1 showed lower maximum stress,strain,and displacement under both material assignment methods,but no clearly suitable scheme was found for the retracted type.Conclusions The outflare type and common type mandibles are more suitable for adopting the osteotomy line design scheme of Scheme 1.For the retracted type,other mandibular angle osteotomy plastic surgery methods may be considered to ensure better biomechanical characteristics.Whether choosing the osteotomy line design scheme or the modeling material assignment method,it is necessary to make the final decision based on the specific analysis objective and resource conditions.

Objective To construct an enhancer-based prognostic risk prediction model for non-small cell lung cancer (NSCLC) by integrating DNA methylome data and transcriptome data. Methods The weighted gene co-expression network analysis (WGCNA) was used to identify NSCLC related genes from the differentially methylated positions (DMPs) of enhancers. Based on the transcriptome data,the prognostic risk prediction model was constructed using LASSO-Cox regression algorithm. Results Through the analysis on DNA methylome data of NSCLC,19784 DMPs were obtained and their distribution patterns were characterized,including 6089 DMPs of enhancers. WGCNA analysis screened 79 highly correlated DMPs of enhancer with NSCLC from the 6089 DMPs. After analyzing the target genes of 79 DMPs with LASSO-Cox regression based on the transcriptome data,10 genes were used to construct a prognostic risk prediction model. The prognostic risk prediction model was evaluated by calculating the areas under the curve (AUC) of 3-,5-,and 10-year time-dependent receiver operating characteristic (ROC) curves in training set and validation set;and the results showed that the 3-,5-,and 10-year AUC in training set and validation set were all higher than 0.7. Finally,a nomogram was constructed to predict the 3-,5-,and 10-year survival probabilities of NSCLC. Conclusion This study provides new insights into the role of enhancers in NSCLC and has the potential to improve the prognosis by guiding personalized treatment decisions.

Objective To conduct preoperative simulations of three different osteotomy line design schemes under centric occlusion based on two distinct material assignment methods,evaluate biomechanical properties of the models,and explore which osteotomy line design schemes are more suitable for different types of mandibles.Methods Three types of mandibles were selected,and CT images were obtained for three-dimensional(3D)reconstruction.Material assignment was completed using the cortical/cancellous bone assignment method and the gray value assignment method.Osteotomy was simulated according to the three osteotomy line design schemes,followed by finite element analysis.Results In all simulation results of the mandibles,the maximum stress was 81.10 MPa,the maximum strain was 0.035 52,and the maximum displacement was 432.4 μm.The stress distributions obtained by the cortical/cancellous bone assignment method showed a larger stress distribution range than that that by the gray value assignment methods,but the maximum stress,strain,and displacement were generally lower.For the outflare type and common type mandibles,Scheme 1 showed lower maximum stress,strain,and displacement under both material assignment methods,but no clearly suitable scheme was found for the retracted type.Conclusions The outflare type and common type mandibles are more suitable for adopting the osteotomy line design scheme of Scheme 1.For the retracted type,other mandibular angle osteotomy plastic surgery methods may be considered to ensure better biomechanical characteristics.Whether choosing the osteotomy line design scheme or the modeling material assignment method,it is necessary to make the final decision based on the specific analysis objective and resource conditions.

Objective To use a physics-informed neural network(PINN)-based model to predict hemodynamics in intracranial aneurysms and address the problems of long simulation time and high computational cost in traditional computational fluid dynamics(CFD)simulations.Methods The PINN model was trained using only the computational domain coordinates and sparse velocity measurement points from CFD data of clinical patients.The predicted blood flow velocity,pressure,and wall shear stress(WSS)from the PINN model were compared with CFD simulation results.Results The proposed method was used to test and validate data from four different patients.For velocity prediction,the average mean absolute error(MAE),average mean relative error(MRE),average mean squared error(MSE)was 4.60％,6.61％,and 0.229％,respectively.For WSS prediction,the average MAE,MRE and MSE was 5.54％,8.58％,and 0.510％,respectively.The PINN model demonstrated a good generalization capability across different aneurysm models and could reduce the computation time of hemodynamics from several hours to just a few seconds.Conclusions The PINN model can effectively compensate for incomplete measurement data through physical constraints,even when boundary conditions are unknown and measurement data are sparse.It can rapidly and accurately simulate the hemodynamics of intracranial aneurysms.This method has the potential to provide effective support for clinical risk prediction in intracranial aneurysms.

Objective To construct an enhancer-based prognostic risk prediction model for non-small cell lung cancer (NSCLC) by integrating DNA methylome data and transcriptome data. Methods The weighted gene co-expression network analysis (WGCNA) was used to identify NSCLC related genes from the differentially methylated positions (DMPs) of enhancers. Based on the transcriptome data,the prognostic risk prediction model was constructed using LASSO-Cox regression algorithm. Results Through the analysis on DNA methylome data of NSCLC,19784 DMPs were obtained and their distribution patterns were characterized,including 6089 DMPs of enhancers. WGCNA analysis screened 79 highly correlated DMPs of enhancer with NSCLC from the 6089 DMPs. After analyzing the target genes of 79 DMPs with LASSO-Cox regression based on the transcriptome data,10 genes were used to construct a prognostic risk prediction model. The prognostic risk prediction model was evaluated by calculating the areas under the curve (AUC) of 3-,5-,and 10-year time-dependent receiver operating characteristic (ROC) curves in training set and validation set;and the results showed that the 3-,5-,and 10-year AUC in training set and validation set were all higher than 0.7. Finally,a nomogram was constructed to predict the 3-,5-,and 10-year survival probabilities of NSCLC. Conclusion This study provides new insights into the role of enhancers in NSCLC and has the potential to improve the prognosis by guiding personalized treatment decisions.

Breast clustered microcalcification (BCM) is one of the most critical X-ray signs of early breast cancer. However, due to the fact that BCM is very tiny and hidden, and manual interpretation is susceptible to subjective factors such as visual fatigue, manual diagnosis of dense BCM based on X-images suffers from a low detection rate, high false-negative rate, and high recall rate. In recent years, with the continuous optimization innovation of deep learning algorithms, and advancements of computer hardware technology, new hope has been brought to accurate diagnosis of BCM, which is expected to realize the accurate assessment of individual breast cancer risk. In this review, the research progress of deep learning for accurate diagnosis of BCM was summarized, such as full-field digital mammography (FFDM), digital breast tomosynthesis (DBT) and contrast-enhanced mammography (CEM), as well as the future developments in this field were discussed.

On August 2-4,2023,the"Third Summit Forum on'Building a Community of Shared Future for Doctors and Patients'"was jointly organized by institutions such as the Chinese Medical Ethics,the Hospital Humanities Management and Talent Training Special Committee of the China Population and Culture Promotion Association,Center for Ethical Studies of Renmin University of China,the Newspaper for China's Physicians,the China Health Law Society,the China Anti-Cancer Association,and the China Association For Ethical Studies in Harbin.The conference arranged a sub-forum for the"Seminar on the Construction of Chinese Medical Humanities",with domestic medical humanities scholars attending the conference.After heated discussions at the seminar,the Scholars'Consensus on the Construction and Development of Chinese Medical Humanities was formed.It was proposed that in the new era,it is urgent to build the medical humanities discipline,as well as lead the academic integration and development of medical humanities under the core socialist values.At the same time,for the construction of the medical humanities discipline,it is necessary to optimize the organizational mechanism,prosper and develop the overall framework of the medical humanities discipline,accelerate the construction of a professional teaching team for the medical humanities discipline,promote the establishment of a new carrier medical humanities education and teaching in cultivating morality and nurturing talents,as well as focus on solving problems related to the cultivation of medical humanities graduate students.

Colorectal cancer is a common malignant tumor of the digestive system globally, with both its incidence and mortality rates increasing annually in China. In recent years, there has been exponential growth in technology based on artificial intelligence, leading to significant advancements in the field of medical imaging diagnosis. Particularly in the application of colonoscopy, CT and magnetic resonance imaging (MRI), artificial intelligence, leveraging its advanced image recognition and feature analysis capabilities, has provided new perspectives for the diagnosis of colorectal cancer, thereby driving the realization of precision medicine. Currently, various artificial intelligence algorithms are either under development or optimization, such as performance comparisons of various artificial intelligence-assisted systems, the collaborative application of multiple algorithms, and integration with other omics. Additionally, challenges persist in the integration difficulty, interpre-tability and credibility, as well as cost and resource limitations of AI in clinical practice, necessitating further standardization and improvement. The authors explore the current status and progress of artificial intelligence in endoscopic and imaging diagnosis of colorectal cancer from four aspects: colonoscopy, CT, MRI and other imaging examination for reference and reference by peers.

Objective To study the effect of plaque eccentricity on stent performance and stress distributions of artery and plaque during stent implantation in stenotic vessels. Methods The stent and idealized stenotic vessels were constructed, and 4 different eccentricities (0%, 20%, 40%, 60%) were attributed to the plaque. Then the stent recoil, stent foreshortening, and stress distributions of artery and plaque when the stent was expanded to the target displacement were analyzed by the finite element method. Results Along with the increase of plaque eccentricity, both stent recoil and stent foreshortening gradually grew. At the same time, the stress of artery and plaque also showed an increasing tread, and high-stress areas gradually approached the narrow side. The maximum von Mises stress of the plaque was much greater than that of the artery. Conclusions Plaque eccentricity had a certain effect on performance of the stent and stress distributions of stenotic vessels. In stent design, the geometry of the plaque should be considered to improve clinical effect of the stent in interventional treatment.